Locomotion Decoding (LocoD): An Open-Source and Modular Platform for Researching Control Algorithms for Lower Limb...

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Signal Processing

Background:

• Current motorized prosthetic legs rely on nonbiological signals, limiting natural control.
• Decoding biological signals (e.g., EMG) from residual limbs offers potential for more intuitive prosthetic control.
• Lack of standardized methods hinders research in decoding algorithms for bioelectric signals.

Purpose of the Study:

• To introduce LocoD, an open-source software platform for unified recording and processing of bioelectric (EMG) and nonbiological sensor data.
• To enable research and benchmarking of control algorithms for prosthetic leg applications.
• To validate the platform's functionality in decoding locomotion modes.

Main Methods:

• Developed LocoD, an open-source platform for EMG and nonbiological signal processing (preprocessing, feature extraction, classification).
• Recorded EMG, IMU, and pressure sensor data from 21 able-bodied participants during various locomotion tasks (walking, stairs, ramps).
• Compared classification accuracy of locomotion modes using three sensor combinations: EMG+IMU+pressure sensor, EMG alone, and IMU+pressure sensor alone.

Main Results:

• The combination of EMG, IMU, and pressure sensors achieved the highest locomotion mode prediction accuracy (93.4% ± 3.9%).
• EMG alone resulted in lower accuracy (74.56% ± 5.8%), while IMU+pressure sensors yielded 90.77% ± 4.6%.
• The superiority of the combined sensor approach was statistically significant (p < 0.001).

Conclusions:

• LocoD is a validated, open-source, and modular platform for researching prosthetic leg control algorithms.
• The platform facilitates the integration and analysis of bioelectric and nonbiological signals.
• Findings demonstrate the enhanced accuracy of combined sensor approaches for prosthetic leg control.